First conference of the BIS CCA Research Network conference on "Incorporating financial stability considerations into central bank policy models"

This conference was hosted by the BIS Representative Office for the Americas in Mexico City. Leading researchers, including Professor E Mendoza (University of Pennsylvania), project academic adviser, provided technical inputs by commenting on the papers and interacting with central bank participants. A central bank roundtable highlighted challenges to the research agenda.

The conference papers focused on models for policy analysis that can capture the feedback between real and financial sectors, by including financial frictions and financial intermediaries. A number of central banks also highlighted country-specific characteristics that allowed responses to questions that are of particular interest to policymakers. In most cases, dynamic stochastic general equilibrium (DSGE) models were used.

The new models can be used to analyse the transmission mechanism or allowed for the analysis of alternative policy instruments (to the monetary policy rate), such as bank capital adequacy ratios, reserve requirements and loan-to-value ratios, to perform counterfactual exercises, and in some cases to forecast macroeconomic outcomes under stress scenarios.

The discussion highlighted some limitations of the current generation of models that incorporate financial stability considerations: (i) in most cases, they cannot be used to analyse how macroprudential instruments correct market failures nor how to design the optimal policy; (ii) they cannot capture nonlinearities we observe during periods of sudden stress (eg sudden increases in spreads) and implications for the interaction between the financial sector and the real economy; (iii) it is not known how well these models describe the economy or financial sector; for example, empirical evidence on the validity of the models is still limited; (iv) many instruments of the existing macroprudential toolkit have not yet been analysed; (v) there is little attention to the role of public debt; (vi) some models are too complex so that transmission mechanisms are difficult to isolate and results are hard to interpret in economic terms; (vii) technical constraints, such as difficulties in solving large nonlinear models with existing algorithms and computing power.